Electronic Letters to:

Endocrine Care: Sara E. Dolan, Jenna R. Kanter, and Steven Grinspoon Longitudinal Analysis of Bone Density in Human Immunodeficiency Virus-Infected Women J Clin Endocrinol Metab 2006; 91: 2938-2945 [Abstract] [Full text] [PDF]

eLetters: Submit a response to this article

Electronic letters published:

Response to the Letter from Drs. Mark J. Bolland and Andrew B. Grey

Steven Grinspoon, Sara E. Dolan, Jenna R. Kanter   (5 February 2007)

Bone density in HIV-infected women

Mark J Bolland, Andrew B Grey   (18 January 2007)

Response to the Letter from Drs. Mark J. Bolland and Andrew B. Grey 5 February 2007
Steven Grinspoon, MD Program in Nutritional Metabolism, Massachusetts General Hospital, Sara E. Dolan, Jenna R. Kanter Send letter to journal: Re: Response to the Letter from Drs. Mark J. Bolland and Andrew B. Grey sgrinspoon{at}partners.org Steven Grinspoon, et al.	We appreciate the opportunity to respond to the thoughtful letter of Bolland et al. in reference to our article, “Longitudinal Analysis of Bone Density in Human Immunodeficiency Virus-Infected Women” (1). Drs. Bolland and Grey raise a number of important issues that have been addressed in the manuscript. The study indeed had a high attrition rate, which we anticipated at the study outset. As we outline in the article, it is difficult to maintain HIV-infected women in 2-year longitudinal, non-interventional, study due to the many psychosocial and sociodemographic barriers to study participation experienced by this population. Nonetheless, we were able to retain sufficient patients to perform longitudinal mixed model analyses with all available data, and show that bone density remains lower over time at the lumbar spine, hip and femoral neck, without differences in the rate of change between the groups in longitudinal follow up. Specific validation analyses were done to demonstrate that subjects who discontinued were not different than subjects that remained in the study, and that the changes in bone density were not different among patients continuing and dropping out of the study. Limiting the modeling to the subjects that finished the study further reduces the power of the study, and does not take advantage of partial, though nonetheless valid data among subjects finishing 6, 12 and 18 months. We specifically obtained bone density at each of these time points to perform longitudinal modeling, and thus we believe our model is the most robust and appropriate for the data. In depth validation analyses were performed to justify this approach. Bolland et al. ask whether the differences in bone density between HIV-infected and non HIV groups remained significant at baseline, controlling for potential confounding factors, including weight, fat mass, smoking and sex steroid levels. The differences between the HIV and non-HIV groups remained significant for the lumbar spine (P=0.02) and for osteocalcin (P=0.04) controlling for these factors. Lastly, as stated in the manuscript, the current study was an extension of our prior study and thus included new longitudinal data for the patients we previously described (2). References 1. Dolan SE, Kanter JR, Grinspoon S. 2006 Longitudinal analysis of bone density in human immunodeficiency virus-infected women. J Clin Endocrinol Metab 91:2938-2945 2. Dolan SE, Huang JS, Killilea KM, Sullivan MP, Aliabadi N, Grinspoon S. 2004. Reduced bone density in HIV-infected women. AIDS 18:475-483
Bone density in HIV-infected women 18 January 2007
Mark J Bolland, Research Fellow Department of Medicine, University of Auckland, Auckland, New Zealand, Andrew B Grey Send letter to journal: Re: Bone density in HIV-infected women m.bolland{at}auckland.ac.nz Mark J Bolland, et al.	The study by Dolan et al. provides potentially important information for HIV-infected women (1). However, interpretation of the results is severely compromised by the high rate of withdrawal during the study. In the HIV group, 47% of participants withdrew by 6 months, 65% by 12 months, and 75% by 18 months, and similar withdrawal rates were observed in the control group. Thus nearly 50% of participants did not have more than one bone density scan. No explanations are provided as to why so many study participants withdrew, nor is there a description of the characteristics of those who failed to complete the study. The authors used mixed-models analysis to compare bone density in the HIV-infected women and the control group over two years of follow-up. All available data were included in these analyses, thus approximately 50% of data were missing. Can the authors provide evidence that these models are stable and valid with such a high rate of missing data? (2) Although the authors reported that the changes in bone density were not different in people who dropped out after 6, 12, or 18 months compared to those who continued beyond these time points, nearly 50% of participants had dropped out by 6 months and therefore do not contribute to these validation analyses. It would be helpful to know that the authors’ models produce similar results when restricted only to those women who had a follow-up bone density measurement and when restricted only to those women who completed the study. In their cross-sectional analysis, the authors reported that bone density was 4-6% lower, and bone turnover higher, in the HIV-infected women compared to the controls. There were differences between the study groups in several variables that might confound the analyses of these skeletal endpoints, namely body weight/fat mass (the HIV group were lighter/had lower fat mass), sex hormone levels (a higher proportion of the HIV group was menopausal or oligomenorrheic), and cigarette smoking (greater exposure in the HIV group). Before concluding that HIV infection causes bone loss or increased bone turnover, the authors should repeat the analyses of BMD and bone turnover after adjustment for these confounders. Finally, it would be helpful if the authors clarified whether the subjects in their previous cross-sectional analysis (3) have also been included in their current analyses. References 1. Dolan SE, Kanter JR, Grinspoon S. 2006 Longitudinal analysis of bone density in human immunodeficiency virus-infected women. J Clin Endocrinol Metab 91:2938-2945 2. Espeland MA, Craven TE, Miller ME, D'Agostino R, Jr. 1999 1996 Remington lecture: modeling multivariate longitudinal data that are incomplete. Ann Epidemiol 9:196-205 3. Dolan SE, Huang JS, Killilea KM, Sullivan MP, Aliabadi N, Grinspoon S. 2004 Reduced bone density in HIV-infected women. AIDS 18:475-483